2x 4TB Seagate SSHD in RAID 1

[davidst99]

Hi, if I 2x 4TB Seagate SSHD in RAID 1 will I loose the speed enhancements these drives provide? Also does it make sense to use these for a non OS drive Thanks.

David

 

[Brado78]

No, the drive wouldn't act any different if you used one. Even the SSD part. The Hybrid is designed to be an OS drive. They don't make good secondary drives, this is coming from a Seagate engineer. (not me)

 

[Victorian Gray]

No, the drive wouldn't act any different if you used one. Even the SSD part. The Hybrid is designed to be an OS drive. They don't make good secondary drives, this is coming from a Seagate engineer. (not me)

fwiw, I use 3 of them (each 4TB) as secondary drives and they work great. All three of them are used to store and serve media (video and music).

 

[davidst99]

Hi, thanks for the info. I'll do some more research.

 

[Data-Medics]

They don't make good secondary drives, this is coming from a Seagate engineer. (not me)

If the Seagate engineer was being totally honest, he'd have told you that Seagate drives aren't good for much of anything....period.

 

[B-Riz]

If the Seagate engineer was being totally honest, he'd have told you that Seagate drives aren't good for much of anything....period.

LOL

I am setting up a monster RAID 0 4 x 3TB Seagate soon, with array ~70% provisioned, should be hella fast.

 

[Data-Medics]

Make sure you back it up nightly. Only a matter of days before one of those fails. Haha.

 

[Data-Medics]

You do know that provisioning is only something you need to worry about on full SSD drives right? Hybrid drives use the SSD portion just as a large cache for the most frequently accessed sectors, and as temporary storage for writes. So the SSD portion will almost always be full no matter how you "provision". And even if that side loses some data due to wear, it's still all stored on the platters. The SSD part is just a second faster copy.

 

[B-Riz]

For me, these are the non hybrid drives.

Technically I am short stroking the drives; so the inner slower tracks are never used.

 

[Insert_Nickname]

I am setting up a monster RAID 0 4 x 3TB Seagate soon, with array ~70% provisioned, should be hella fast.

Unless you're actually needing all that space, a single 250GB SSD is going to walk all over that RAID array.

This is coming from someone who used to RAID 10.000RPM Raptors back in the day. Then got an X25-M, never looked back at HDDs for system drives...

Technically I am short stroking the drives; so the inner slower tracks are never used.

Last time I checked, it said 2015 in the calender. Short stroking drives is so '90s...

 

[B-Riz]

Quote:
Originally Posted by B-Riz
I am setting up a monster RAID 0 4 x 3TB Seagate soon, with array ~70% provisioned, should be hella fast.
Unless you're actually needing all that space, a single 250GB SSD is going to walk all over that RAID array.

This is coming from someone who used to RAID 10.000RPM Raptors back in the day. Then got an X25-M, never looked back at HDDs for system drives...

Quote:
Originally Posted by B-Riz
Technically I am short stroking the drives; so the inner slower tracks are never used.
Last time I checked, it said 2015 in the calender. Short stroking drives is so '90s...

LOL

I have this growing Steam library, and have collected the 3TB drives over the past year from sales and what not.

So, I figured why not?

 

[B-Riz]

This is coming from someone who used to RAID 10.000RPM Raptors back in the day. Then got an X25-M, never looked back at HDDs for system drives...

I really wanted to get 4 x 1TB Raptors, but, they are stuck at ~$200 a piece, and these 3TB Seagates are very quick in RAID 0, been happy with the 2 x 3TB setup right now.

 

[bunnyfubbles]

If the Seagate engineer was being totally honest, he'd have told you that Seagate drives aren't good for much of anything....period.

quit being lazy

https://www.backblaze.com/blog/best-hard-drive-q4-2014/

Which Hard Drive Should I Buy?

All hard drives will eventually fail, but based on our environment if you are looking for good drive at a good value, it’s hard to beat the current crop of 4 TB drives from HGST and Seagate. As we get more data on the 6 TB drives, we’ll let you know.

 

[Data-Medics]

Did you actually read the chart where it shows Seagate models with annual failure rates in the 20-43% range? While HGST is in the 0.4-3% range.

Granted the 3Tb are the worst culprets, but the 4Tb are the same architecture. They just haven't had as much time to age yet.

And the OP I responded to said he planned to use 4, 3Tb in a RAID 0. Do the math:
(4 Drives) x (43% annual failure rate) = 172% statistical chance the array will fail in the first year alone. Not good odds.

 

[Essence_of_War]

(4 Drives) x (43% annual failure rate) = 172% statistical chance the array will fail in the first year alone. Not good odds

This is a highly dubious understanding of how statistics work, you cannot possibly have a greater than 100% chance of array failure, nor can you have a less than 0% chance of array survival. By the same token, you can't have a greater than 100% chance of array survival, nor a less than 0% chance of array failure.

Anyway, the probability of at least 1 drive failure is 1 - the probability of 0 drive failures, and assuming identical independent random variables (and yes, I know that this assumption isn't great, but it is a starting point and probably a lower bound):

1 - .57^4 = 89%

Which, you'll note, is considerably less than 172%, and considerably more sensible.

It's easy to want there to be a "buy X brand for reliability" answer, but real life is harder than that. There are particular models that are better than others, but generally brands don't uniformly produce either good or bad products.. Backblaze's results certainly indicate that you should stay away from the particular 1.5 TB and 3TB models from Seagate that they tested, but the 4TB models appear to have quite a different annualized failure rate, I find the claim that they just haven't had them running for long enough and that somehow they will regress to a 5-10x higher annualized failure rate sometime soon to be highly dubious. I find it even more dubious when there isn't large sample size data backing up the claim.

 

[B-Riz]

I have 2 x 4TB Seagate's to use for regular backup, in one of these: http://www.mediasonic.ca/product.php?id=1357290977 (very happy with it)

Both the 3 and 4 TB drives are rated for the same power on hours, but the 4TB is 5400 RPM and the 3TB is 7200 RPM.

Also, the 4TB has lower power requirements, which I would guess helps with longevity.

http://www.seagate.com/www-content/...am/en-us/docs/desktop-hdd-ds1770-5-1409us.pdf

The Backblaze high failures are from shucked drives.

I imagine, when digging into the SMART raw data they provide, those failed 3TB's are outside the mfg spec of power on hours and load / un-load cycles (which I want to explore once I am done semi-upgrading main desktop)

 

[B-Riz]

Did you actually read the chart where it shows Seagate models with annual failure rates in the 20-43% range? While HGST is in the 0.4-3% range.

Granted the 3Tb are the worst culprets, but the 4Tb are the same architecture. They just haven't had as much time to age yet.

And the OP I responded to said he planned to use 4, 3Tb in a RAID 0. Do the math:
(4 Drives) x (43% annual failure rate) = 172% statistical chance the array will fail in the first year alone. Not good odds.

FWIW, I have been waiting the past year for one of the 3TB's to fail, but the SMART data reports are clean, and no corruption / failure has occurred.

So, on to the grand 4 x 3TB RAID 0 experiment!